active 2025-10-19 → 2026-06-28 (UTC)
Activity over time
Daily event counts in the loaded window
Line chart, 253 days from 2025-10-19 to 2026-06-28. Pushes: 17 total, peak 8 in a day. Pull requests: 0 total, peak 0 in a day. Issues: 23 total, peak 23 in a day. Comments: 6 total, peak 6 in a day. Stars: 0 total, peak 0 in a day.
- Pushes
- Pull requests
- Issues
- Comments
- Stars
Stars, PRs, issues and forks are under-captured in the later part of this window. GH Archive progressively stopped capturing non-push events during 2026 — −95% or worse by the end of the window. Every series here except Pushes fades for that reason, so a decline above reflects the archive, not this repository. Pushes stay reliable throughout, so read them, and the contributor counts derived from them, as the real signal. Data health has the measurements.
Top contributors
Pushes, PRs, issues, reviews and comments — stars and forks excluded, so this is contribution rather than popularity
| Contributor | Contributions | Pushes | PRs | Comments |
|---|---|---|---|---|
| agorevski | 46 | 17 | 0 | 6 |
Recent activity
Latest issues, pull requests and releases
- Issue#101agorevski2026-06-27 06:09[Evaluation] Sample eval-limit rows reproducibly instead of taking the first N
- Issue#100agorevski2026-06-27 06:09[Inference UX] Make web inference configuration docs accurate and complete
- Issue#86agorevski2026-06-27 06:08[Training Efficiency] Reuse existing dataset splits instead of re-splitting every run
- Issue#118agorevski2026-06-27 05:06[Data Quality] Add a versioned schema for decontamination-critical metadata
- Issue#109agorevski2026-06-27 05:06[Observability] Surface inference trace paths and token diagnostics in web results
- Issue#105agorevski2026-06-27 05:04[Evaluation] Add first-class baseline comparison to the eval CLI
- Issue#84agorevski2026-06-27 05:03[Training Efficiency] Expose quantization/precision mode in train.py CLI
- Issue#73agorevski2026-06-27 04:41[Inference/Web] Surface model readiness and effective config before decompile jobs
- Issue comment#74agorevski2026-06-27 04:41[Inference/Web] Expose validated generation controls in the web/API decompile flow
- Issue#58agorevski2026-06-27 04:00[Training] Batch evaluation generation during training runs
- Issue#42agorevski2026-06-27 03:47[Pipeline Review] Standardize validation split filename across pipelines
- Issue comment#42agorevski2026-06-27 03:47[Pipeline Review] Standardize validation split filename across pipelines
- Issue comment#41agorevski2026-06-27 03:47[Pipeline Review] Use tokenizer-based sequence length detection in training wrappers
- Issue#38agorevski2026-06-27 03:47[Pipeline Review] Make DeepSpeed precision match GPU capability
- Issue comment#31agorevski2026-06-27 03:47[Pipeline Review] Group dataset splits to prevent contract/function leakage
- Issue#55agorevski2026-06-27 03:16[Pipeline Review] Decode ABI custom errors in REVERT TAC metadata
- Issue#45agorevski2026-06-27 03:15[Pipeline Review] Do not mark HuggingFace contracts processed after silent compile failures
- Issue comment#45agorevski2026-06-27 03:15[Pipeline Review] Do not mark HuggingFace contracts processed after silent compile failures
- Issue#45agorevski2026-06-27 03:14[Pipeline Review] Do not mark HuggingFace contracts processed after silent compile failures
- Issue#40agorevski2026-06-27 03:14[Pipeline Review] Seed evaluation sampling for reproducible metrics
- Issue#39agorevski2026-06-27 03:14[Pipeline Review] Derive inference TAC budget from configured context length
- Issue#25agorevski2026-06-27 03:14[Pipeline Review] Parse failures are reported as high-confidence safe results
- Issue#24agorevski2026-06-27 03:14[Pipeline Review] Bytecode integer-overflow fragments are never reported
- Issue#10agorevski2026-06-27 02:18Medium: Training loss is computed over prompt/input tokens
- Issue#11agorevski2026-06-27 02:18Medium: Dataset split ratios are misleading
Totals cover only the window loaded into ClickHouse and count events, not GitHub's lifetime totals — 0 stars here means stars gained during the window, not the repo's star count.